In this paper we developed a vehicle routing problem (VRP) model. We applied this
model for determining the distribution?s routes of urea fertilizer from a depot to retailers. In daily
distribution, each truck can serve more than one route, i.e., multiple trips, and each retailer has
a time window. Additionally, we modeled the vehicle routing as a mixed linear programming
(MILP) to minimize the total transportation cost. Moreover, we used the road network on a
digital map instead of calculating Euclidean distance of distributor to retailers and inter-retailer
ones. Our approach is more realistic than the classical one. Finally, we test the model using the
historical distribution data, and the model can minimized the total cost about 2.28% compared to
the original routes.